A rescaling algorithm for multi-relaxation-time lattice Boltzmann method towards turbulent flows with complex configurations

Abstract

Understanding and modeling flows over porous layers are of great industrial significance. To accurately solve the turbulent multi-scale flows on complex configurations, a rescaling algorithm designed for turbulent flows with the Chapman-Enskog analysis is proposed. The mesh layout and the detailed rescaling procedure are also introduced. Direct numerical simulations (DNSs) for a turbulent channel flow and a porous walled turbulent channel flow are performed with the three-dimensional nineteen-velocity (D3Q19) multiple-relaxation-time (MRT) lattice Boltzmann method (LBM) to validate the accuracy, adaptability, and computational performance of the present rescaling algorithm. The results, which are consistent with the previous DNS studies based on the finite difference method and the LBM, demonstrate that the present method can maintain the continuity of the macro values across the grid interface and is able to adapt to complex geometries. The reasonable time consumption of the rescaling procedure shows that the present method can accurately calculate various turbulent flows with multi-scale and complex configurations while maintaining high computational efficiency.